Loss-based Objective and Penalizing Priors for Model Selection Problems

Changwoo J. Lee

Many Bayesian model selection problems, such as variable selection or cluster analysis, start by setting prior model probabilities on a structured model space. Based on a chosen lo…